Objectifs
- Reply in minutes (not 1–2 days) to stop losing high-intent buyers on social channels
- Scale ~2,000 daily conversations across WhatsApp, Instagram and Messenger without growing headcount
- Lift conversion on inbound DMs by selling, qualifying, and closing inside the chat itself
Solutions
- Centralized WhatsApp, Instagram and Facebook Messenger conversations on Unifunl
- Added a website chat widget on BrazilianGlow to capture and convert site visitors as a new funnel
- Deployed a multilingual AI digital employee (Derja, French, Arabic) to recommend products and handle objections
- Used Unifunl AI Agents to qualify buyers, process orders, and close deals 24/7. Humans only step in on edge cases
Brazilian Glow is a fast-growing cosmetics brand based in Tunisia, selling across WhatsApp, Instagram and Facebook Messenger. With customers expecting near-instant answers and most conversations happening inside DMs, chat is the front door of the business. As inbound volume grew to around 2,000 conversations per day, scaling response quality without scaling headcount became the priority. So Brazilian Glow partnered with Unifunl to unify those channels and add a website chat widget that turns site visitors into another converting funnel.
The problem: manual replies couldn’t keep up with social commerce demand
Brazilian Glow’s acquisition engine was working. Meta ads and organic content were funneling huge volumes of curious shoppers into Instagram, WhatsApp and Messenger. The bottleneck wasn’t demand. It was speed.
Across DMs, Messenger and WhatsApp, the support team was juggling around 2,000 conversations per day, each one fragmented across a different inbox. Without a unified view, context was lost between channels and agents repeated the same answers dozens of times an hour.
On top of that, the brand’s own website was generating real traffic but had no live touchpoint. Visitors who hesitated in front of dozens of SKUs simply bounced, with no way to ask a question in the moment. It was a constant leak in the funnel.
After 5pm, and across the weekend, the pressure broke down completely. Customers asked about shades, stock and shipping at the exact moment they were ready to buy, then waited 1–2 days for a response. In social commerce, that’s not a delay. It’s a lost order.
The Unifunl solution
To fix the bottleneck without inflating headcount, Brazilian Glow rebuilt its inbound flow around a single omnichannel inbox powered by Unifunl. WhatsApp, Instagram and Facebook Messenger were unified into one conversation layer, so every message lands in the same place with full customer context, regardless of the channel it came in on.
On top of that inbox, Brazilian Glow deployed a Unifunl AI digital employee. It isn’t a basic FAQ bot. It’s an autonomous agent trained on the full product catalogue, pricing, shades, shipping rules and objection patterns. It runs the entire pre-sale workflow end to end, in Tunisian Derja, French and Arabic.
Reduce response times with AI Agents on every channel
The AI Agent now handles the high-frequency tasks that used to consume the support team: greeting incoming DMs, asking the right qualifying questions, recommending products in real time and keeping the conversation moving without delay.
Average first response time dropped from 1–2 days to under 2 minutes, even at 2am and even on Sundays. That single change unlocked an entirely new layer of revenue: shoppers who would previously have gone cold are now answered while they’re still in buying mode.
Open a brand new converting funnel with the website chat widget
Brazilian Glow’s website was generating strong traffic, but visitors had no live touchpoint. Dozens of SKUs, no way to ask a question in the moment, and a single bounce killed the sale. As part of the rollout, Unifunl added a website chat widget on brazilianglow.com.tn powered by the same AI agent.
Instead of leaving shoppers to read product pages on their own, the widget greets visitors, asks a few short questions about skin type, hair type, concerns and goals, and recommends the exact product (or bundle) that matches. It works like a great in-store advisor.
The result was a brand new converting funnel. Visitors who used to bounce now move from “I don’t know what to pick” to “add to cart” inside a single chat, turning the website from a passive catalogue into a live sales channel that runs alongside DMs.
Recommend products, handle objections and close orders automatically
Instead of agents copy-pasting answers, the Unifunl agent interprets each shopper’s intent, suggests the right product or shade, and addresses objections (price, delivery time, suitability) with on-brand responses tested for conversion.
When a customer is ready to buy, the agent collects order details, confirms shipping and books the order directly inside the chat, whether the conversation started on Instagram, WhatsApp, Messenger or the website widget. Human agents only step in for genuine edge cases, which now represent less than 1% of conversations.
“I am no longer thinking about how we’re going to hire new salespeople. Revenue is steady, we can see and control everything, and we keep making sales 24/7, even after hours.”
The results
By unifying its DMs and adding a brand new website chat funnel, Brazilian Glow turned every touchpoint into a 24/7 sales engine. First response time dropped from 1–2 days to under 2 minutes. That’s a 10x acceleration on the moment that matters most in social commerce, across roughly 2,000 conversations per day.
Conversion on inbound conversations climbed to 40% across DMs and the new website widget, and total order volume grew by 30% within months. Growth was driven both by orders captured outside working hours and by website visitors who finally got the product guidance they needed to buy with confidence.
Most importantly, Brazilian Glow scaled all of this without growing the support team. Agent headcount was reduced by 70%, the AI now resolves the vast majority of conversations end-to-end, and the team is free to focus on launches, content and high-value customers instead of typing the same reply for the hundredth time.
